Regional, national, and international comparison of baseline characteristics and comorbidity from MI patient registries

Registry populations of MI patients differ considerably according to the inclusion and exclusion criteria of different series. Differences in inclusion criteria comprise the time period from symptom onset, patient age, possible exclusion of cardiogenic shock,

unstable patients and early death, heterogeneous application of varying definitions of MI over the years, and inclusion of patients other than type I MI. These differences have to

69 be kept in mind when the baseline characteristics and comorbidity of various MI

registries are compared. The following considerations on baseline patient data refer to several registries with a similar population background (i.e. urban population of

developed western world countries), namely the Berlin Myocardial Infarction Registry (BMIR), the Italian BLITZ registry, the American National Registry of Myocardial Infarction (NRMI), and the Euro Heart Survey.

The Italian BLITZ registry collected data of almost 2,000 MI patients from 296 coronary care units (87% of all active Italian units at the time of the study), who were enrolled according to previous MI definitions (2000 and earlier) and up to 48 h after symptom onset over a two week period in October 2001.¹⁰⁸ Hence this study reflects a broad and likely to be almost representative Italian MI population, but may differ from our current study due to different inclusion criteria, and due to the short and remote inclusion time period.¹⁰⁹ Due to fact that the patients included in our study were at the same time part of the BHIR database, the Berlin MI Registry was the natural reference for this

investigation. The United States NRMI was the largest and longest running MI registry in the United States, collecting data from 1990 to 2006 of over 1.3 million patients with acute MI and a short symptomatic interval (<12 h after symptom onset) from 2,157 hospitals.¹¹⁰ The Euro Heart Survey ACS-III enrolled 19205 MI cases from 138 hospitals and 21 countries over a 2 year inclusion period from 2006 to 2008.¹¹¹ The proportion of STEMI patients out of the overall MI population was 50.7% in our study. The STEMI proportion varies between the different registries, mainly according to the inclusion criteria, but in part also over time. The NRMI, with the very short symptomatic interval of enrolled patients of no longer than 12 h, comprised 54.6% of STEMI patients. It is worth noting, however; that over the years, the STEMI proportion within the NRMI series changed from over two-thirds to one-third.¹¹²

The BMIR also enrolled a much higher proportion of STEMI patients in its early

inclusion period until 2002 (around 73%), while in 2003 to 2004 the proportion of STEMI decreased to 62%, and the proportion of NSTEMI showed a respective increase from around 23% to 38% of enrolled acute MI cases with a short symptomatic interval. The change in STEMI/NSTEMI proportions, on the one hand, reflected a change in

definitions of MI and may, on the other hand, also be due to the enhanced recognition of NSTEMI through the broader application of cardiac troponin testing. The Italian BLITZ registry also enrolled patients according to early MI definitions and reported a similarly

70 high STEMI proportion¹¹³ , whereas the Euro Heart Survey reported a STEMI proportion of approximately 60% (presumed new left bundle branch block patients included)¹¹⁴, or an even lower proportion of around 45% in earlier studies.¹¹⁵ Taken together, the STEMI proportion in this study is within the expected range in a registry enrolling from 2008 to 2011.

The mean age of acute MI patients in this study was 66.6 y, which is in line with data from the BLITZ registry (67)¹¹⁶. The mean age of STEMI patients was somewhat lower (64 y), a finding comparing well to data from the Euro Heart Survey ACS III registry (running from 2006 to 2008 and including ACS patients within <12h from symptom onset) (64 y)¹¹⁷ and the BMIR. A similar mean age was documented by the American NRMI registry, where the mean age varied from 64 y to 66 y) and in the BLITZ registry (66 y for STEMI patients). It is a well-known fact that STEMI patients are approximately 2 y younger than NSTEMI patients. In addition, a huge difference in age is usually present between men and women suffering from an acute MI, with female patients having their MI 7 to 9 y later compared to male patients. In our study, the difference in mean MI manifestation age between women and men was 8.4 y, again within the expected range for a current MI population. Gender distribution in our study revealed the proportion of women at 30%. This proportion was similar for all MIs and STEMI cases. Comparable gender distribution data for MI patients are available from the BLITZ registry (female proportion in STEMI 29%, for all MI 30%).¹¹⁸ The Euro Heart Survey ACS III registry also reports similar data with 28-29% of STEMI patients being

female.¹¹⁹ The early BHIR analysis from 1999 to 2004 showed varying and overall somewhat higher proportions of female MI cases during these years (33.8% to 35.7%

for STEMI, and 34.7% to 42.2% for NSTEMI).¹²⁰ The gender distribution of our study is compatible with the assumption that our population reflects a routine MI and STEMI cohort.

The chief symptom in our overall study population was chest pain in the vast majority of patients (81.9%). This is a common finding, although the extent of the preponderance of chest pain in MI varies. While the number of patients with chest pain in the BLITZ registry was higher at 92.6%¹²¹ , other sources like a population based series from the Worcester Heart Attack study¹²² and a series from Australia¹²³ reported a lower

prevalence of chest pain. In the Worcester study, chest pain was shown to be age-dependent and sex-age-dependent, and the prevalence of chest pain varied between 46%

71 and 81%, whereas the Australian MI series reported an overall incidence of this chief complaint in 78% of patients. Taken together, our study population appears to be well in line with these previous series concerning symptomatic presentation.

The time of presentation and admission is known to impact on the care of MI patients and the outcome, in the sense that presentation during off-hours is linked to longer DTB times and higher mortality.¹²⁴ Out of our study population, 35% of all MI patients and 36% of STEMI patients were admitted during working hours. The proportion of STEMI patients admitted during working hours is slightly lower in our study compared to the proportion reported by an earlier analysis (2004 to 2007) from the BMIR, where the off-hour admission rate of STEMI cases in Berlin amounted to 38.9%.¹²⁵ In the American NRMI registry, the proportion of STEMI/LBBB patients presenting during regular

working hours was 37%.¹²⁶ Out of the 7,655 STEMI patients included in the Euro Heart Survey ACS III registry, 42% were admitted during working hours.¹²⁷ In the ACTION-GWTG database study (2007-2010) involving 43,242 patients with STEMI, 37% of the patients presented during working hours.¹²⁸

Cardiovascular risk factors and comorbidity

In order to make the comparative presentation of cardiovascular risk factors and

comorbidities clearer and more transparent, we confronted the data of the present study with the data of the above mentioned large registries in two tables (see section 6

“Supplementary Tables”, Table 11) for overall MI cohorts and a second one for STEMI patients (Table 12). Compared to the BLITZ registry, the Euro Heart Survey and the overall BMIR series our study shows a numerically higher burden of comorbidity with a higher prevalence of diabetes, cigarette smoking, a higher rate of renal failure, heart failure and a higher prevalence of prior coronary revascularization (PCI and/or/CABG), although the pattern of comorbidity still appears to be compatible with previous studies.

The observed differences, in particular the difference compared to the overall BMIR data, suggest that the studied hospital population appeared to be particularly comorbid;

which may be due to the fact that Berlin is a heterogeneous city of socio-economic and ethnic diversity and a city with inhomogeneous age distribution between different districts. The present study just provides insight into a segment of mainly three Berlin districts that are covered by the single hospital for this study. The higher rate of prior coronary revascularization may reflect the more recent study period of the present investigation and the probably better access to this treatment option.

72 The pattern of comorbidity may also be influenced by administrative factors. In Germany the reimbursement system for hospitals changed from orientation to length of hospital stay to a diagnosis related group (DRG) system. The then new system enforced more accurate documentation of the principal diagnosis and cost-enhancing comorbidities.¹²⁹ Accordingly, the incidence of hypertension, hypercholesterolemia and smoking

displayed a rapid increase between 1999 and 2006.¹³⁰

The observed higher prevalence of smoking in our study is probably the result of

regional differences within the city of Berlin. A second fact that has to be kept in mind is that we included former smokers in addition to current smokers into this parameter. This could also contribute to the difference compared to the BLITZ and Euro Heart Survey registries.

Altogether, despite the limited size of our study and the above mentioned particularities, the distribution of baseline characteristics, risk factors, and comorbidities in our study was in the expected range compared to other registries. This suggests that our findings are likely to be reliable and to some extent also generalizable.

4.3 Main categorical variables